Two-Dimensional Boron - ACS Publications - American Chemical

to produce a contour plot in which the x and y axes are IlB chemical shift (at 115.8 MHz) and the z direction is peak intensity. Peaks on the diagonal...
1 downloads 0 Views 978KB Size
J. Am. Chem. S o t . 1984, 106, 29-31

29

Two-Dimensional Boron- 11-Boron- 11 Nuclear Magnetic Resonance Spectroscopy as a Probe of Polyhedral Structure: Application to Boron Hydrides, Carboranes, Metallaboranes, and Metallacarboranes' T. Leon Venable, William C. Hutton, and Russell N. Grimes* Contribution f r o m the Department of Chemistry, University of Virginia, Charlottesuille, Virginia 22901. Received June 13, 1983

Abstract: A general technique has been developed for the direct determination of boron-boron atom connectivities in all types of boron cage compounds, based on J-correlated, two-dimensional (2-D) I'B-IIB N M R spectroscopy. In this method, which is the first reported demonstration of homonuclear 2-D N M R for a quadrupolar nucleus, a Jeener-type sequence is employed to produce a contour plot in which the x and y axes are IlB chemical shift (at 115.8 MHz) and the z direction is peak intensity. Peaks on the diagonal ( F , = F2) correspond to the "normal" one-dimensional "B N M R signals, while the off-diagonal (cross) peaks at 6,, bB arise from spin-spin coupling between IlB nuclei A and B. Since the observable scalar coupling is limited to short-range (adjacent-atom) interactions, the cross peaks reveal the boron connectivity pattern in the framework and routinely yield information about the distribution of electron density that is most difficult to obtain otherwise. Assignment of l'B resonances in conventional 1-D spectra is also greatly facilitated, in many cases allowing a complete assignment in a single experiment usually requiring 10-30 min including generation of the contour plots. Although some discussion of theory is included, this paper is directed to the utilization of the method by the synthetic chemist and describes its application to a variety of boron cage species. A general observation (with two known exceptions) is that boron atoms linked by bridging hydrogen atoms usually do not exhibit mutual coupling, in agreement with earlier theoretical predictions that the electron density in B-H-B bridge bonds is negligible on the B-B vector. A step-by-step procedure for deducing framework structures from 2-D spectra is presented in an appendix.

Since its discovery in 1946,2 nuclear magnetic resonance spectroscopy has played an increasingly important role in organic, and more recently inorganic, chemistry. Its power as a tool for the study of molecular structure has grown steadily with the introduction of stronger magnetic fields, heteronuclear and multiple-frequency experiments and spin d e ~ o u p l i n gand, , ~ in very recent years, two-dimensional (2-D) N M R . 4 In the broad area of boron cluster chemistry, N M R utilizing the IlB nucleus has been particularly important in the characterization of polyhedral cage structure^;^ the rapid development of this field in the last 20 years could scarcely have occurred in its absence. Yet there are serious limitations in "B N M R as a structural probe for polyhedral clusters, and they are becoming increasingly conspicuous as more complex borane and heteroborane species are synthesized. The two primary problems, well-known to boron chemists, are (1) broad signals (50-150 Hz) arising in part from unresolved I1B-I1B coupling6 and (2) the absence of a broadly applicable theory of "B shifts. The first of these difficulties, which leads to frequent overlap of resonances, is only partly overcome in high-field (>4 T ) instruments; coupling information is normally restricted to first-order boron-terminal hydrogen interactions, as "B-"B coupling is seldom observable (even with the aid of line-narrowing technique^,^ IIB-IIB coupling is still not resolved ( I ) Presented in part at the Symposium on New Structural Techniques in Inorganic Chemistry, National Meeting of the American Chemical Society, Seattle, WA, March 1983, Abstract INOR-39. (2) (a) Purcell, E. M.; Torrey, H. C.; Pound, R. V. Phys. Rea. 1946, 69, 37. (b) Bloch, F.; Hansen, W. W.; Packard, M. E. Ibid. 1946, 69, 127. (3) (a) Baldeschwieler, J. D.; Randall, E. W. Chem. Rea. 1963,63, 81. (b) Hoffman, R. A,; Forsen, S. Nucl. Magn. Reson. 1966, 1 , 15. (4) (a) Jeener, J., Ampere International Summer School, Basko Polje, Yugoslavia, unpublished results, 1971. (b) Bax, A,; Freeman, R.; Morris, G. J . Magn. Reson. 1981, 42, 164. (c) Bain, A. D.; Bell, R . A,; Everett, J. R.; Hughes, D. W. Can. J . Chem. 1980,58, 1947. (d) Nagayama, K.; Wiithrich, K.; Ernst, R. R. Biochem. Biophys. Res. Commun. 1979, 90,305. ( e ) Kumar, A.; Wagner, G.; Ernst, R. R.; Wiithrich, K. Ibid. 1980, 96, 1156. (f)Wagner, G.; Kumar, A,; Wiithrich, K. Eur. J . Biochem. 1981, 114, 375. (5) (a) Eaton, G. R.; Lipscomb, W. N. "NMR Studies of Boron Hydrides and Related Compounds"; Benjamin: New York, 1969. (b) Todd, L. J.; Siedle, A. R. Prog. Nucl. Magn. Reson. Spectrosc. 1979, 13, 87. (c) Siedle, A. R. Annu. Rep. N M R Spectrosc. 1982, 12, 177. (6) Weiss, R.; Grimes, R. N. J . Am. Chem. Soc. 1978, 100, 1401.

0002-7863/84/ 1506-0029$01.50/0

in most cases8). The presence of the InB isotope ( I = 3, 20% abundance) introduces a further complication via coupling with the observed IlB nucleus. The second problem is intrinsic to boron cage frameworks. Unlike the chemical shifts exhibited by 'H and I3C nuclei, "B ( I = 3 / 2 , 80% abundance) exhibits few general, reliable correlations with molecular structure. The "B shifts are largely determined by the paramagnetic term uo in Ramsey's equation' and can vary considerably with even minor variations in structure. Consequently, it is frequently the case that the "B N M R spectrum of a boron cluster of unknown geometry conveys little more than symmetry information, Le., the presence or absence of mirror planes or rotation axes. When superposition of inequivalent peaks occurs, even this can be misleading. A further perspective on this problem is given by the fact that, even in molecules of established geometry such as BInH,,, the complete unambiguous assignment of IlB N M R spectra may require laborious studies involving isotopic labeling or single-frequency decoupling. I n Confronted with these difficulties, we became intrigued with the possibilities afforded by the recently developed 2-D N M R technique^,^^" which involve the detection of a frequency response over two independent time domains. In a 2-D experiment, the data are collected as a t , , t2 data matrix followed by a double Fourier transform to yield a frequency spectrum that reveals the presence of connectivity patterns (in this case, J-coupling connectivities) between various nuclei. The first 2-D N M R experiments involving the IlB nucleus utilized the heteronuclear lH-"B interaction and revealed the correlations between the I I B reso(7) Clouse, A. 0.;Moody, D. C.; Rietz, R . R.; Roseberry, T.; Schaeffer, R. J . Am. Chem. Soc. 1973, 95, 2496. (8) Line narrowing in "B NMR cannot resolve couplings in which J < -10 Hz or in which the coupled nucleus, S, relaxes so quickly that the observed nucleus *seesn only an average spin state of S. (9) Ramsey, N. F. Phys. Rea. 1950, 78, 699; 1953, 90, 232. (10) The complete assignment of the spectrum of B,,H,, via isotopic labeling experiments over a period of several years is described in ref 5a, pp 164-170. ( 1 1) (a) Aue, W. P.; Bartholdi, E.; Ernst, R. R. J . Chem. Phys. 1976,64, 2229. (b) Macura, S.; Huang, Y.; Suter, D.; Ernst, R. R. J . Magn. Reson. 1981, 43, 259.

0 1984 American Chemical Society

30 J. A m . Chem. Soc., Vol. 106, No. 1 , 1984 nances and the directly bound protons.I2 This approach is analogous to those involving spin nuclei such as I3C and 31P.13 The homonuclear experiment has been successfully applied to the spin nuclei IH, 13C, and 31P,all ofwhich typically exhibit resolvable spin-spin coupling in the conventional ( 1- D ) spectrum.l4 It was not obvious that the homonuclear 2-D experiment involving IlB would be feasible. Not only is I’B-llB coupling normally unresolved in the one-dimensional spectrum but also relaxation times ( T I )for I’B nuclei are 100-1000 times shorter (milliseconds) than are those for protons, suggesting that relaxation would average the spin-coupling interaction between spin pairs to zero. However, we have successfully applied the Jeener 2-D procedure (to be described) to homonuclear lB-l IB systems, representing the first such extension of this technique to a quadrupolar nucleus. Following our earlier short c o m m ~ n i c a t i o n we l ~ have refined the method, confirmed its general applicability, and applied it to a variety of boron cage species. In this paper, which is oriented toward the practical utility of the method for the synthetic/ structural chemist, we do not deal in depth with the implications of this work for the theory of magnetization transfer through scalar connected energy levels in quadrupolar nuclei. However, qualitative comments on these aspects will be given.

Experimental Section The compounds utilized in this study were synthesized in this laboratory by previously published procedures or purchased commercially and used as received. Samples of 1,7-C2BioH12(m-carborane) and B I o H i 4 [decaborane(l4)] were obtained from Alfa Inorganics, and 1,2-C,Bi,H,, (o-carborane) was purchased from the Dexsil Chemical Corp., Hamden, C T 06514. The spectrum of nido-2,3-(C2Hs)2C2B4H616 was obtained on a neat sample in a sealed tube, and that of Na+(C2Hs),C2B4H,- was recorded in T H F solution sealed under vacuum. Tetrahydrofuran (THF) was dried over LiAlH, and distilled in vacuo. All other compounds are air-stable materials and were dissolved in an appropriate solvent in 10mm tubes. Solvents were dried prior to use. Spectra were recorded unlocked in a stable (measured drift rate < 2 X H 2 / h ) magnetic field (8.5 T ) with a “B resonance frequency of 115.8 M H z . Referencing was to an external solution of BF3.0Et2. Uninterrupted, incoherent ‘H decoupling at low levels (-2 W) was employed. The C O S Y pulse sequence (vide infra) was used to generate the tl, t 2 data matrix in which t l was incremented by the inverse of the sweep width in the F , dimension and r2 has the same meaning as the acquisition time in a 1-D experiment. Typically, the tl, t 2 matrix was collected as 128 X 256 (reals plus complex) data points. The data were zero-filled once in each dimension after apodization by using a cosine bell function. The final F , , F2 matrix was symmetrized” prior to plotting the absolute value spectra. The lower limit for the detection of a reliable cross peak on our spectrometer is -1% of the height of the tallest diagonal peak. Spectra were obtained on a Nicolet Magnetics Corp. NT-360/0xford spectrometer, and all data manipulation utilized standard Nicolet software with a 1280/293B data system. Workup of the data including contour plots usually required 10-20 min. By use of the Nicolet M A C R O capability the collection and processing of data, including a first contour plot, have been automated for routine analysis.

Results and Discussion Description of the Technique. The two-dimensional spectra generated by this experiment contain two types of peaks, those (12) (a) Finster, D. C.; Hutton, W. C.; Grimes, R. N. J . A m . Chem. Sot. 1980, 102, 400. (b) Colquhoun, I. J.; McFarlane, W. J . Chem. SOC.,Dalton Trans. 1981, 2014. (1 3) (a) Bodenhausen, G.; Freeman, R. J . Mugn. Reson. 1977, 28, 471. (b) Freeman, R.; Bodenhausen, J . Am. Chem. Sot. 1978, 100, 320. (c) Morris, G. A.; Hall, L. D. J . Am. Chem. Sot. 1981, 103, 4703. (14) (a) Bax, A,; Freeman, R. J , Magn. Reson. 1981, 44, 542. (b) Freeman, R.; Morris, G. J . Magn. Reson. 1981, 42, 164. (c) Lynn, D. G.; Phillips, N. J.; Hutton, W. C.; Shabanowitz, J.; Fennell, D. I.; Cole, R. J . J . Am. Chem. Sot. 1982, 104, 7319. (d) Van Divender, J. M.; Hutton, W. C. J . Mugn. Reson. 1982, 48, 272. (15) Venable, T.L.; Hutton, W . C.; Grimes, R. X . J . Am. Chem. SOC. 1982, 104, 4716. (16) Hosmane, N. S.; Grimes, R. N. Inorg. Chem. 1979, 18, 3294. (17) Baumann, R.; Wider, G.;Ernst, R. R.; Wiithrich, K . J . Mugn. Reson. 1981, 44, 402.

Venable, Hutton. and Grimes lying along the diagonal ( F , = F2) and the off-diagonal or “cross peaks”. The plots are symmetrical about the diagonal so that the cross peaks are found a t 61, 6, and equivalently a t S,, 6 , . The off-diagonal peaks reveal spin-spin coupling interactions between adjacent boron nuclei and hence disclose the atom connectivities in the polyhedral framework. These coupling patterns are normally clearly observable even when, as is usually the case, no IlB-l’B coupling is evident in the routine 1-D spectrum. W e emphasize that the method utilizes standard FT N M R equipment together with 2-D software available from the instrument manufacturer or developed by the user. The pulse sequence COSY (correlated spectroscopy) involves 16 cycles and may be written

where T, is a time sufficient to ensure that the system is in thermal equilibrium. The first ( ~ / 2 pulse ) creates transverse magnetization as in a routine 1-D experiment, and the spins are allowed to process for a time t l so that they become frequency labeled. The time t l is incremented by the inverse of the sweep width for the first frequency domain Fl in order to build up a t , , t2 data array; at the end of tl a second ( a / 2 ) pulse is applied. This mixing pulse transfers magnetization between coupled spins so that the observed spin now reflects a modulation based on the chemical shift difference (6, - 6,) of the two scalar coupled spins. At the end of this pulse the transverse magnetization is acquired for t 2 , as in any 1-D spectrum. A qualitatively useful analogy for this experiment is to consider each resonance line in a 1-D spectrum as a receiver that is capable of detecting the resonance frequencies of spins that share J-coupling energy levels with the “receiver”. The receiver resonance is phase-modulated by the coupled spins as the evolution time is incremented by t l . The phase shifts +2 and +, for the second pulse and receiver, respectively, have the dual purpose of suppressing axial peaks and permitting quadrature detection in both dimensions.I8 Extending this cycle by phase shifting the first pulse cancels artifacts that arise from quadrature detection errors.18 The time domain data matrix S(tl,t2) that is collected is then double Fourier transformed to give a frequency domain spectrum S(FI,F 2 ) . The resulting spectra can be presented in “stacked” fashion as shown in Figure la, but these are time-consuming to produce and are less readily analyzed, although they can be useful in revealing the relative intensities of cross peaks and diagonal peaks. More generally useful, for our purposes, are absolute value contour plots (e.g., Figure lb) in which the x and y axes represent “ B chemical shifts [measured relative to BF3.O(C2H5), with positive values downfield]. Resonance intensities are represented in the z direction (normal to the page) as a series of contours. The peaks along the diagonal represent transitions of spins whose precession frequencies were not modulated during the experiment and the chemical shift information they contain is identical with that obtained by normal (1-D) N M R . The cross peaks arise from spin-spin IIB-IIB coupling, which for boroncontaining clusters is significant only for adjacent nuclei. The connectivities revealed by the cross peaks therefore correspond directly to the framework arrangement of boron atoms within the cluster. W e turn now to a discussion of specific spectra chosen to illustrate both the power of the method in structure elucidation and spectral assignment, as well as its limitations. Spectra of Boron Clusters: BloH14. Figure 1 presents the proton-decoupled 2-D IIB-IIB spectrum of BloHI4 with the conventional I’B(lH] 1-D spectrum below. The four peaks along the diagonal correspond to the resonances in the 1-D spectrum and from the pattern of cross peaks it can be seen that the 5 , 7 , 8, and 10 boron atoms (identified from the unique area-4 signal) are coupled to two of the three other types of b o r o n atoms. The resonance a t 6 -36 ppm exhibits cross peaks with all other reso(18) (a) Bax, A. ”Two-Dimensional Nuclear Magnetic Resonance in Liquids”; Delft University Press: Delft, Holland, 1982. (b) Hoult, D. 1.; Richards, R. E. Proc. R. Sot. London, Ser. A 1975, 344, 3 1 1 ,

J . Am. Chem. SOC.,Vol. 106, No. 1 , 1984 31

2 - 0 "B-"B N M R Spectroscopy

I I " " / " " l " " I- '"O " l " "- 2 I0

-30

-qO

PPM

~,

10

C

,

-10 ,

, ,

-20 ,

, I ,

,I?

-30 , ,

,

I

-YO

PPM

-Y@

PPM

L

l ' " ' 1 " ' ' l ' " / ' " " ' ' ' ' i

IO

0

-IO

-20

-30

-YO

PPM

IO

0

-IO

-20

-30

Figure 1. Two-dimensional proton-decoupled IIB-IIB FT NMR spectra of B,,H,, in toluene: (a) stacked plot; (b) contour plot. In these and subsequent figures. the conventional one-dimensional IlB soectrurn. shown at the bottom, correlates with the peaks on the diagonal. All spectra were recorded aF25 0C except where otherwise indicated.

nances and can therefore be assigned to B(2, 4). It follows that the lowest field peak must arise from B( 1, 3) and the remaining resonance from B(6, 9). A complete assignment of the BloH14 llB spectrumlo is thereby achieved in a single experiment. Even more significantly, the cage structure can be determined from the 2-D NMR data independently of the crystallographically established geometry,I9 if standard electron-counting rules for polyhedra20 are assumed to hold (see the Appendix). No cross peak appears between the B(5, 7 , 8, 10) and B(6, 9) signals, reflecting a general observation to which we have found only two exceptions thus far: coupling between boron nuclei linked by bridging hydrogen atoms is usually absent within the limits of detection of the experiment. This finding is consistent with theoretical calculations2'a* on boron hydrides, which show that electron density in B-H-B bridges is negligible along the boron-boron axis,21dsince the scalar coupling that results in the appearance of a cross peak is dependent on the electron density between the two nuclei (the exceptions mentioned will be discussed later in the paper). BloH,42-.The spectrum of the decaborane( 14) dianion, measured as the dicesium salt in aqueous solution, is shown in Figure the 2, 4 resonance is identified from 2. As in the case of BIOHL4 its coupling with all other boron sites in the molecule, but here it lies a t the low-field rather than the high-field end of the spectrum. The 5, 7 , 8 , 10 resonance can be assigned from its 1-D integration, and while the remaining two diagonal peaks exhibit

rH

';I

l-z

Figure 2. Spectrum of ( ~ C S ) ~ + ( B , ~inHH~2~0)at~ 40 (19) (a) Tippe, A,; Hamilton, W. C. Inorg. Chem. 1969, 8, 464. (b) Moore, E. B.; Dickerson, R. E.; Lipscomb, W. N. J . Chem. Phys. 1957, 27, 209. (c) Kasper, J. S.; Lucht, C. M.; Harker, D. Acta Crystallogr. 1950, 3, 436. (20) (a) Wade, K. Adu. Inorg. Chem. Radiochem. 1976, 18, 1. (b) Rudolph, R. W. Acc. Chem. Res. 1976, 9, 446. (c) Mingos, D. M. P. Nature (London), Phys. Sci. 1972, 236, 99. (21) (a) Switkes, E.; Stevens, R. M.; Lipscomb, W. N. J . Chem. Phys. 1969,51,2085. (b) Switkes, E.; Epstein, I. R.; Tossell, J. A,; Stevens, R. M.; Lipscomb, W. N. J . Am. Chem. Sac. 1970, 92, 3837. (c) Laws, E. A,; Stevens, R. M.; Lipscomb, W. N . Ibid. 1972, 94, 4467. (d) Evidence for this effect was previously noted in 1-D I1B spectra of 6-CH,BloHl,: Sprecher, R. F.; Carter, J. C. Ibid. 1973, 95, 2369.

OC.

similar cross peak patterns, the one near 6 - 36 ppm is a triplet in the proton-coupled 1-D spectrum22(not shown) and hence must arise from the BH2 groups at the 6, 9 positions. The presence of a cross peak between the B(5, 7 , 8, 10) and B(6, 9) resonances (in contrast to B10H14)is as expected, since these boron-boron connections are not hydrogen bridged in this molecule. (22) Lipscomb, W. N.; Wiersema, R. J.; Hawthorne, M. F. Inorg. Chem. 1972, 11, 651.

32 J. Am. Chem. SOC.,Vol. 106, No. 1, 1984

Figure 3. Spectrum of

Venable, Hutton, and Grimes

l,2-C2B,oH12 in n-hexane.

Figure 5. Spectrum of

,

nido-2,3-(C2H5),C2B,H6 (neat).

%

L

0

c,t

R .

Figure 4. Spectrum of

'

r/"

l,7-C2BloH12 in n-hexane.

J

'

Figure 6. Spectrum of

. -., n Na+[2,3-(C,H5)2C2B4H5]in THF 30

i/

l,2-C2BloH,2and 1,7-C2BloH,2( 0 - and m-Carborane). The spectra of the two isomers are presented in Figures 3 and 4, respectively. In each case the 1-D spectrum, shown below, exhibits four signals in a 2:2:4:2 area ratio, consistent with the established structures that have CzLpoint group symmetry; however, in the 1-D spectra23only the area-4 peak can be unambiguously assigned. Moreover, from the I-D spectra alone it is not possible to differentiate between the two isomers. Analysis of the 2-D spectra, however, clearly identifies the isomers and (with one assumption) permits assignment of all resonances in the spectrum of 1,2C2BIOHIz. Thus, only the 1, 2 isomer contains boron atoms (8, 10) that are directly linked to all other boron sites; only in the 1 , 7 isomer is there a boron environment (2, 3) that is adjacent to only one other boron type (4,6,8, 11). Both features are clearly

exhibited by the cross peak correlations in these spectra, so that in each case two of the four boron resonances can immediately be assigned. In the spectrum of the 1, 2 isomer, an ambiguity remains between the lowest and highest field peaks, in that the connectivity indicated by the cross peaks is the same for each resonance. A distinction can be made, however, if one assumes that the 3, 6 boron nuclei, each adjacent to two carbon atoms and hence relatively shielded, will resonate a t high field. For 1,7-C2B,0H12the identification of the B(5, 12) and B(9, 10) resonances is not possible from these data alone; the assignments shown were obtained elsewhere from isotopically labeled derivativesz4 Thus, in molecules of high symmetry such as the

(23) Vickers, G. D.; Agahigian, H.; Pier, E. A.; Schroeder, H.Inorg. Chem. 1966, 5 , 693.

(24) Garber, A . R.; Bodner, G . M.; Todd, L. J.; Siedle, A. R. J . Magn. Reson. 1917, 28, 383.

J . A m . Chem. Soc., Vol. 106, No. I , 1984 33

2 - 0 “B-”B N M R Spectroscopy

-. i

i

_-

..--

+/

“-

\“”+~%-*-h-vII”

---

23

-. b

y

~--

Figure 8. Spectrum of w(4,S)-HgC1[ 1,2,3-(C5[CH,] ,)Co(CH3),C2B3H,].

Figure 7. Spectrum of ~ , F ’ - H ~ ( B , Hin~ benzene. )~

icosahedral carboranes, even the 2-D method may not permit complete or unambiguous identification of peaks. In such cases, a simple solution would often be to lower the symmetry by substitution (e.g., by alkylation a t one of the carbon vertexes). I I ~ ~ O - ~ , ~ - ( C , H , ) ~ Cand ~ B[2,3-(C2Hs)2C2B4HsT. ,H~ The 2-D spectrum of the neutral carborane (Figure 5) shows the expected coupling between the apex boron B(l) and the basal boron atoms B(5) and B(4, 6), but coupling between B(4, 6) and B(5) is absent-another manifestation of the bridge hydrogen effect already noted in the BlOHl4spectrum. On bridge deprotonation with N a H to form the m ~ n o a n i o n(Figure ~~ 6), a B(4, 6)-B(5) cross peak appears, reflecting the presence of electron density along the B-B vector. (Tautomerism of the remaining bridge proton between two equivalent locations retains the mirror symmetry of the parent molecule on the NMR time scale.) The spectrum of the anion contains extraneous peaks arising from decomposition products, but in the 2-D spectrum these are clearly distinguishable from the resonances of (C2H5)2C2B,Hs-by virtue of the absence of intermolecular cross peaks. This illustrates another general advantage of 2-D spectroscopy: the ability to resolve the spectra of mixtures into their individual components. In our experience this particular feature has been of considerable value in a number of cases. and p( 4,5) -HgCI[ 1,2,.3-C5( C H 3 )&oM,p’-Hg( BsH,) (CH,)2C2B,H4]. The 2-D spectra of these mercury-bridged complexes26(Figures 7 and 8) suggest that the heteroatom bridge does not perturb the electron distribution within the cage. No cross peaks are observed for the cobaltacarborane cage, which contains both a B-H-B and a B-Hg-B bridge, implying that the absence of coupling in B-H-B groups may extend to boronmetal-boron bridges as well. Except for the apical B(1, 1’) resonance in the spectrum of Hg(B,H8),, unique assignments cannot be made without additional information. ~~~o-~,~,~-[C,(CH,)~]F~(C~H,)~C,B,H,. This ferra~arborane~’ can be described as a derivative of r z i d 0 - 2 , 3 - ( C ~ H ~ ) ~ C , B(vide ~H~ supra) in which the apical B H unit is replaced by an isolobal Fe[C6(CH,),] group or, equivalently, as a sandwich complex of C6(CH3)6, Fe2+, and the cyclic (C2H5)2C2B3H,2-ligand (an electronic analogue of CsHS-). Its 2-D spectrum (Figure 9) clearly contains a cross peak, indicating coupling between the hydro(25) Onak, T.; Dunks, G . B. Inorg. Chem. 1966, 5 , 439. (26) (a) Hosmane, N. S.; Grimes, R. N. Inorg. Chem. 1979, 18, 2886. (b) Finster, D. C.; Grimes, R. N. Inorg. Chem. 1981, 20, 863. (27) Swisher, R. G.; Sinn, E.; Grimes, R. N. Organomefallics 1983, 2, 506.

I’ ,

J

25

2c

C

E

OPP

Figure 9. Spectrum of nrdo-1 ,2,3-[C6(CH3),]Fe(C2H,),C2B3H5i n nhexane.

gen-bridged boron nuclei. This observation runs counter to the cases discussed above involving B-H-B and B-Hg-B bridged species and indicates that there are important differences in the basal (ring) bonding in the iron complex in comparison to 2,3(C2H,),C2B4H6and other B-X-B bridged compounds. Possibly, the delocalization of electron density about the C2B3ring is greater in the iron complex (resembling its isoelectronic C,H,- counterpart) so that the B-H-B bridges in this compound approach “protonated double bond” character; in the R,C2B4H6carborane framework28 the bonding in the B-H-B groups may well be more localized. While such questions are obviously not resolved by this study, 2-D “B-llB NMR clearly has considerable potential for probing the electron distribution in boron cages and can reveal distinct differences in the bonding of related molecules that may not be apparent from X-ray crystallographic studies. 2-(C2HS)CoB4H8. The 2-D spectrum of this small metall a b ~ r a n e(Figure ~~ lo), a B5H9analogue, provides the only other ~

(28) Boer, F. P.; Streib, W. E.; Lipscomb, W. N. Inorg. Chem. 1964, 3, 1666.

34 J . Am. Chem. SOC.,Vol. 106, No. 1 , 1984

Venable, Hutton, and Grimes

'

I

/

I

I

r

L1

-El

i - . , , . , . , , , , , . , , , , , , , , , , IO 5 0 -5

. , , ,

, , ,

, ,

-20

-15

-IC

, ,

~

-25

PP*

/I

A

.iIL

I\ z& 3.c

8

s

r

-5

-1c

-20

-25

I

11

8

-15

il

I

'M' .'it

PPP

:I_

-

Figure 10. Spectrum of 2-(CsH5)CoB4H8in heptane.

I v_

Figure 12. Spectrum of 2-(C,H5)CoB,H,,

Figure 11. Spectrum of 6-[CS(CH3)S]CoB,H,3in toluene

example we have encountered of detectable I1B-l1B coupling between hydrogen-bridged nuclei: a weak, but real, cross peak is observed between the B(3, 5 and B(4) resonances. A rationale for this observation is not readily apparent although one theoretical d e s ~ r i p t i o nof~ ~2-(C5H,)CoB4H8suggests that the replacement of BH in B5H9 by Co(C,H,) leads to enhanced boron-boron bonding in the cluster; however, other aspects of that particular study have been challenged r e ~ e n t l y . ~ It ' should also be noted that the crystallographically determined3* B(4)-B(3,5) bond length [ 1.757 (2) A] is slightly shorter than the corresponding bond length in B5H, (1.77 and 1.803 A from X-ray33aand midata, respectively). The weak intensity of the B(3,5)-B(4) cross peak suggests correspondingly weak boron(29) Miller, V. R.; Weiss, R.; Grimes, R. N. J . Am. Chem. SOC.1977, 99, 5646. (30) Brint, P.; Spalding, T. R. J . Chem. SOC.,Dalton Trans. 1980, 1236. (31) DeKock, R. L.; Fehlner, T. P. Polyhedron 1982, I , 521. (32) Sneddon, L. G.; Voet. D. J . Chem. SOC.,Chem. Commun. 1976, 118. (33) (a) Dulmage, W. J.; Lipscomb, W. N. Acta Crystallogr. 1952, 5, 260. (b) Hrostowski. H. J.; Myers, R. J. J . Chem. Phys. 1959, 22, 262. (c) Schwoch, D.; Burg, A. B.; Beaudet, R. A. Inorg. Chem. 1977, 16, 3219.

--

in n-hexane.

0

LiO

l

~

"

l

20

"

'

l 0

'

"

"

- 20

'

~

"

' ~ -YO FPM

LiO

20

0

-?O

-qO

PPM

Figure 13. Spectrum of 5-[C,(CH,)S]CoB,H,3 in n-hexane

boron coupling, but such interpretations must be taken with caution (vide infra). (C5R5)CoB9HI3 Isomers (R = H or CH3). Four isomeric cobaltaboranes can be formally generated from the BI0Hl4structure (Figure 1) by the replacement of a BH unit with an isoelectronic Co(C5H5)or Co[C5(CH3),]moiety, and three of these, the 6-,

2- D

' I B-I'

B N M R Spectroscopy

J. Am. Chem. Soc., Vol. 106, No. 1 , 1984 35

Table I compound

6-CI-5.7-1C , K H J j C o ] 2B,H,,34e in toluene

(CH,),C,B,H,

38

Sa

(assignment)

4 9 . 9 [ B(6)] 24.4 [ B(1, 3 ) ] 20.1 [ B(2)] 7.9 [ B(8, 1 O)] -6.1 [ B ( 9 ) ] -39.8 [ B ( 4 ) ]

in acetone

v l / * - Hzb

149 95 129 175 146 78 53 134

-3.9 -7.4 -11.0 -41.6 ~~~

~~

iB(10)] [ B(1)] [B(11)] IB(5)l

72 53 33 59

T , , nisC 6.9 5.9

5.1 3.5 2.5 9.6 31.6 -28 39.6 28.3 26.6 86.3

"natural linc width", Hzd 46 54 62 91 127 33 10 11

8.0 11

12 3.7

~~

aShifts in ppm relative to BF,.O(C,H,), with positive shifts downfield. bhleasurcd from a least-squares fit of a Lorentzian curve to the observed resonance. T , values measured by inversion-recovery using composite ( n T , ) - ' . e Insufficiently resolved for clear measurements of u l i z and T I .

2-, and 5-isomers, have been prepared and characterized by X-ray c r y ~ t a l l o g r a p h y(the ~ ~ 1-isomer has yet to be reported). Figures 11-1 3 depict the 2-D l i B spectra of these corn pound^^^ and are particularly illustrative of the value of the 2-D method in structure determination, since in all three cases the coupling patterns lead to a unique geometry and provide a complete, unambiguous assignment of the spectrum. The only a priori assumptions are (1) a decaborane-like (nido) framework consistent with the skeletal electron count and (2) the absence of coupling between B-H-B bridged nuclei, in accordance with the empirical rule already noted. In the 6- and 2-isomers, a mirror plane is evident from the 2:2.2:1:1:1 pattern of the 1-D spectra, and of the three area-1 peaks in each spectrum (arising from borons on the mirror), one is distinguished by having only one cross peak with other resonances; it is therefore assigned to the 9-vertex, which is directly (Le., other than B-H-B) linked to only one other boron, B(4), whose resonance in both spectra is thereby identified. Although B(4) is linked to two pairs of boron atoms [B(8, 10) and B(1, 3)], in the 6-isomer these sites are distinguishable by their cross peaks since B(8, 10) is not linked to any unique (area-1) boron other than B(4); B(1, 3) is connected to two unique boron atoms, B(4) and B(2). Thus, the identities of the B(2) and, finally, the B(5, 7) signals fall into place. The case of the 2-isomer is analyzed similarly (its geometry can also be inferred from the fact that only'the 2- and 6-isomers generate a 2:2:2:1:1:1 1-D spectrum). The B(6) resonance has no cross peaks, since it is linked to other borons only via hydrogen bridges. In this spectrum the B(5, 7) and B(8, 10) diagonal peaks are so close that the cross peak, if present, is obscured; however, the B(8, 10) signal is identified from its coupling to the unique boron B(4). The 5-isomer is the only possible BloH14-like(C,R,)CoB,H,, structure lacking symmetry and on this basis was identified from its 1-D N M R spectrum29prior to X-ray c h a r a ~ t e r i z a t i o n .Given ~~~ only the 1-D spectrum, none of the resonances can be assigned, but the 2-D information (Figure 13) readily identifies all nine peaks (see the Appendix). Again, a n expected cross peak [B(7)-B(8)] cannot be seen owing to overlap of the two signals along the diagonal. A systematic approach to the analysis of 2-D IIB-liB spectra is given in the Appendix. Further General Observations. The transitions responsible for cross peaks that appear in 2-D spectra cannot be entirely explained by the classical model of precessing and an analysis of the interactions between nuclei is facilitated by a density matrix approach.Ii The contour map is directly related to this mathe(34) (a) Pipal, J. R.; Grimes, R. N. Inorg. Chem. 1977, 16, 3251. (b) Zimmerman, G. J.; Hall, L. W.; Sneddon, L. G. Ihid. 1980, 19, 3642. (c) Wilczynski, R.; Sneddon, L. G. Ihid. 1979, 18, 864. (d) Gromek, J. M.; Donohue, J. Cryst. Struct. Commun. 1981, 10, 871. (e) Venable, T. L.; Sinn, E.; Grimes, R. N. Inorg. Chem. 1982, 21, 895. (35) The "B spectra of the (CSH,)CoB9H13complexes and their CS(CHI), counterparts are essentially identical.

matical treatment and is, in fact, a pictorial representation of the matrix that describes the J spin-spin coupling network for the observed nuclei. Our concern here is not with these quantitative aspects but with the structural features implied by the presence of a cross peak. A cross peak will appear when four conditions are met: ( 1 ) sufficient electron density exists between the nuclei to permit scalar coupling; (2) the two nuclei are not decoupled by a longitudinal relaxation time ( T i )that is short relative to the reciprocal of the coupling constant (Le., 2xJTIs